Anthropogenic climate change is perhaps the most significant existential challenge that humanity faces today (Mora et al., 2018; Gills and Morgan, 2020). A rapid increase in utilization of fossil fuels since the industrial revolution has increased the concentration of greenhouse gases (GHGs) in the atmosphere at a faster rate than has been observed previously (Peng et al., 1983; Etheridge et al., 1996; Lacis et al., 2010). The overwhelming majority of scientific evidence points to this increase in atmospheric GHGs, specifically carbon dioxide, being the cause of the changing global climate (Oreskes, 2004; Hartmann et al., 2013). Historically, there has been an equilibrium between CO₂ sequestration via photosynthesis and CO₂ emissions by biodegradation and other natural mechanisms that gradually removed CO₂ from the atmosphere, transforming Earth's atmosphere into the habitable one that we now rely on (Des Marais, 2000; Gonzalez Hernandez and Sheehan, 2020). Burning fossil fuels to power today's society introduces a new, rapid flux of CO₂ into the atmosphere that natural photosynthesis can no longer compensate (Grace, 2004; Le Quéré et al., 2018).

Renewable technologies must be developed and deployed to bring equilibrium back to the global carbon cycle (Holdren et al., 1980; Goreau, 1990; Gielen et al., 2019). There are several renewable approaches to CO₂ emissions mitigation, and ultimately the planet will require a diversified portfolio as no single technology could reasonably abate global CO₂ emissions alone (Moriarty and Honnery, 2012; Fasihi et al., 2019; Realmonte et al., 2019). Many viable solutions rely on the increased utilization of renewable (wind, solar, hydroelectric, and others) energy with low lifecycle CO₂ emissions intensity (Sims, 2004). Displacing fossil fuel-based electricity generation with renewables indirectly reduces GHG emissions by preventing them from being emitted in the first place. Carbon capture and storage (CCS) technologies, on the other hand, provide direct CO₂ emissions reduction by sequestering CO₂ from anthropogenic sources or from the air (Snæbjörnsdóttir et al., 2020). Lastly, carbon capture and utilization (CCU) affords both direct and indirect emissions reduction by directly utilizing captured CO₂ as a reactant to make a product, which in turn displaces the product's fossil-derived counterpart. Together, the latter two technologies are known as carbon capture, utilization, and storage (CCUS). CCUS is a subset of negative emissions technologies (NETs), which is defined for the purposes of this article as technologies that reduce atmospheric GHG concentrations below the concentration that would occur without the technology (McLaren, 2012).

The utility of NETs increases over time as we implement measures to reduce global GHG emissions. In today's world where coal is still burned for electricity, it is in many cases more advantageous to replace these most polluting emitters with renewable energy to prevent their CO₂ emissions in the first place (Gaffney et al., 2020). Once the “lowest hanging fruit” is taken by removing the worst emitters, the relative difficulty of mitigating CO₂ increases. At this intermediate point, the utility of NETs to capture CO₂ from clean emitters is comparable to deploying additional indirect CO₂ emissions reduction measures. Ultimately, when most electricity generation is done renewably (which introduces additional energy storage challenges that further enable several CCU technologies) there will still be substantial CO₂ emissions in several areas, such as the chemical and aerospace industries, agriculture, aviation, and cement production (Fasihi et al., 2019). At this point, NETs will be required to maintain equilibrium in the global carbon cycle. As of 2020, we need to remove more than 170 gigatons of CO₂ from the atmosphere to remain under the target of 1.5°C of warming by 2100 (Johansson et al., 2020). This is the equivalent of more than 5 times the total amount of CO₂ emitted globally in 2019 (Friedlingstein et al., 2020).

Currently, the most widely commercialized systems that capture CO₂ recover it from a high-purity source, such as hydrogen production or sugar fermentation (de Assis Filho et al., 2013). These CO₂ recovery (CR) systems utilize feed streams with typically >95% CO₂ concentration, requiring minimal impurity removal to increase the concentration to >99%, while compressing to provide a liquefied CO₂ product for ease of transport and use. In contrast, point source CO₂ capture (PSC) uses a more dilute feed stream, such as natural gas flue gas which is ~4–16% CO₂ in N₂ and O₂ from the air (Jiang et al., 2019). A testament to the energetic implications of CO₂ concentration on NET efficiency, much PSC research focuses on source gases with higher CO₂ concentrations (Li et al., 2011; Baker et al., 2017). Lower concentration feed streams necessitate use of a higher volume sorption system, and typically uses the industry standard monoethanolamine (MEA) CO₂ scrubber. Direct air capture (DAC), on the other hand, captures CO₂ from the air at ~416 ppm which is much more energetically challenging than either CR or PSC (Ren et al., 2021). A key differentiator between these technologies is the concentration of CO₂ in their source material, which dictates the feedstock mass and energy required for separation.

In this article, we present an analysis of CO₂ capture technologies as they interface with downstream continuous flow CO₂ utilization systems. We provide an industrial perspective by discussing the advantages and challenges to deploying flow reactors downstream from CR at high-purity point sources in the Air Company pilot reactor in New York, as compared to PSC from the flue gas of a natural gas-fired power plant using an amine absorption-based system at the Air Company demonstration reactor in Calgary. In doing so, we highlight the trajectory for deployment of CO₂ capture technologies more broadly, when coupled with utilization as these emerging fields and respective technologies scale. Lastly, we discuss the lessons learned from these deployments as they relate to integrated CO₂ capture and utilization systems and their associated technological and infrastructural barriers.

All three major categories of CO₂ capture technologies are subject to the same development constraints as other chemical processes. To progress from conceptual idea, to a proof of concept, followed by a benchtop prototype, then pilot, demonstration, and small commercial plant is highly capital intensive and requires research and development infrastructure of its own. As they are both still the subject of heavy R&D, the economics for both PSC and DAC do not reach the low cost point that CR has achieved. Given the long chemical scale-up and development cycles and the fundamentally higher energy requirement for PSC and DAC, CR is the lowest hanging fruit today as a CO₂ source to provide suitable feedstock for CO₂ utilization. On the other hand, CR has the lowest potential for scale and long-term CO₂ emissions reduction since there are limited sources; many are dependent on industries powered by fossil fuels that can be replaced with renewable alternatives, such as H₂ production (Table 1).

CR acts on CO₂-rich gases produced during processes such as fermentation and is fully commercially available (Haszeldine et al., 2018). Due to its source gas containing >90% CO₂, it can function without high thermal energy input to capture and release CO₂ from a sorbent and is typically fully powered by electricity. This gives it the best economics of all the CO₂ capture technologies today, as evidenced by its widespread use to produce CO₂ for sale. Due to its high concentration feedstock, CR is the most entropically and energetically favorable CO₂ capture technology by a large margin as shown in Table 1. However, because CR has the lowest potential for scale for global CO₂ removal, it is insufficient to meet decarbonization goals over the long term if not used in tandem with PSC and DAC. Due in part to its low potential for scale and need for future technological improvements, there is a dearth of scientific literature and policy-based focus on CR. For this reason, it makes the most pragmatic sense to prioritize deployment of CR in locations where there are concentrated CO₂ streams being emitted as the lowest-hanging fruit in the NET portfolio. These deployments can be done rapidly while simultaneously continuing to scale PSC and DAC technologies.

PSC predominantly acts on post-combustion point sources, such as natural gas combined cycle (NGCC) plants with CO₂ concentrations of 4–16% (Jiang et al., 2019; see Table 1). Currently, PSC is operating at industrial scales, but its poor economics prevent widespread deployment, prompting further R&D in laboratories and pilot plants to reduce capital and energy costs (Table 2). Several PSC pilots show significant promise to further these goals, with innovations such as corrosion inhibitors helping to reduce heat duty from 5.0 to 1.8 GJ/ton CO₂ and with projected return on investment within 2.5 years (Idem et al., 2015; Shirmohammadi et al., 2018). Current challenges are centered around optimizing adsorbent capacity at the high temperatures and low CO₂ concentrations present in flue gas streams (Divekar et al., 2020). Unfortunately, many otherwise promising improvements in PSC (including selective membranes in combined cycles reaching 90% capture rates) require expensive retrofitting of plants, which may deter commercial deployment (Turi et al., 2017).

PSC typically uses the byproduct gas of fossil fuel combustion as a source of CO₂, raising concerns that overreliance on PSC may enable continued fossil fuel dependence. In contrast, DAC has been proposed as a mechanism that minimizes the need for infrastructural change or fossil fuel dependence from its inception (Lackner et al., 1999). As DAC removes CO₂ from air, with a concentration that is orders of magnitude lower than that of CR and PSC, it requires the highest energy input of the technologies studied (Breyer et al., 2019). One of its strengths as a component of a diversified portfolio of NETs is in offsetting distributed emissions, such as those from aviation and agriculture. Areas for improvement of DAC technology primarily focus on decreasing energy requirements for CO₂ desorption (see Table 3). Of the high temperature and low temperature DAC systems, low temperature thus far achieves lower heat supply costs (Fasihi et al., 2019). Thicker absorbent films, thinner monolithic walls, and adsorbents with higher efficiency at ambient air conditions can decrease the required energy of temperature vacuum swing adsorption (Sinha et al., 2017).

Critics argue that both PSC and DAC delay an inevitable transition to renewables, and thus increase the societal costs associated with pollution (Jacobson, 2019, 2020). There are scenarios in which this concern is understandable; some DAC deployment trajectories require a quarter of global energy demand by the end of the century (Realmonte et al., 2019). However, the warming targets in the Paris Agreement can only be met if NETs are part of the portfolio of climate solutions deployed (Haszeldine et al., 2018), making it imperative to deploy both low-carbon energy generation and NETs. The timing and trajectory of NET deployment is critical to reconcile both sides of the discussion. In the near term, NETs have the most impact by displacing the processes that are both most CO₂ intensive, thereby maximizing both direct and indirect CO₂ emissions, and have no renewable replacement in the foreseeable future.

We propose a trajectory for deployment of CO₂ capture technologies that follows this approach, and examples on the pilot and commercial demonstration scales using continuous flow CO₂ conversion systems are described below. The stranded sources of concentrated CO₂ are urgent to capture, but not as important in the long-term as our ability to remove CO₂ from the air. Correspondingly, the energy and capital intensity of DAC make it more economic to first pursue lower-hanging fruit, especially when today's renewable energy infrastructure does not provide adequate heat to power DAC systems without burning fossil fuels (Holmes et al., 2013; Keith et al., 2018), but it is critical that large-scale R&D efforts lower these expenses in the future. Proper timing for both renewable electricity and NETs is critical to maximize societal benefit.

Air Company's CO₂-to-alcohols commercial pilot plant in Brooklyn, New York is a distillery that utilizes CO₂ delivered from CR-sourced sites, including fermentation facilities, to produce ethanol. The ethanol is distilled, mixed, and bottled on-site to produce spirits and hand sanitizer. The facility is powered by a mixture of offsite wind turbines and utility-scale solar photovoltaics, enabling ethanol production from CO₂, H₂O, and renewable electricity. Figure 1A is a photograph of Air Company's CO₂ conversion pilot plant in the facility, with a NEL H-series H₂O electrolysis system and a fixed bed flow reactor for CO₂ hydrogenation. A flow chart of the process can be seen in Figure 1B; in brief, CO₂ and H₂ are compressed, heated, and fed into the 16-foot fixed bed flow reactor. The reactor is filled with a novel and proprietary heterogeneous catalyst that has not yet been reported in literature and is developed and synthesized on the kg-scale in Air Company's facilities, enabling stable and continuous conversion of CO₂ and H₂ into ethanol. The gaseous products are passed through a condenser assembly, which separates the room-temperature crude ethanol aqueous liquid and gases (Sarp et al., 2021). Room temperature gases are recycled into the reactor for further conversion. The facility and systems are further adaptable to accommodate CO₂ electrolysis when that technology is at an appropriate commercial stage of development (Chen et al., 2018).

After the CO₂ conversion process, the crude ethanol is filtered and distilled to produce a neutral spirit that is ~95% ethanol by volume, the remaining 5% being water with <300 ppm net of all impurities by gas chromatography (GC), and meets all requirements for United States Pharmacopeia (USP) grade. In this facility, the CO₂ fed into the reactor is captured offsite via CR powered by renewables, which typically requires 120 kWh/ton, together with ~30 kWh/ton for transportation. CR is a continuous flow system that delivers CO₂ as a liquid at high pressure, ensuring consistent supply and eliminating concern about the rate or variability of CO₂ use in flow reactors downstream.

Defining NETs as technologies that reduce atmospheric GHG concentrations below that which would occur without the technology, lifecycle assessment is based on a cradle-to-gate analysis (McLaren, 2012). Under optimal production conditions and operating at capacity, to produce 1 kg of ethanol as a functional unit, a minimum of 1.91 kg of CO₂ and 0.26 kg of H₂ is required. Depending on the lifecycle analysis methodology, a reasonable carbon footprint for the captured CO₂ is ~ −1.78 kgCO₂e given that the CO₂ used in the process would otherwise be emitted to the atmosphere (Müller et al., 2020). The primarily wind and solar power for the facility has average lifecycle emissions of 10 gCO₂e/kWh (Sovacool, 2008). Water electrolysis consumes 81 kWh/kg H₂, which equates to 0.21 kgCO₂e (NEL Hydrogen, 2020). Compression, heating, cooling, and distillation are all powered by electricity or waste heat and net ~8 kWh/kg ethanol. Amortized material of construction emissions over production lifetime averages 40 gCO₂e/kg ethanol, accounting for unoptimized system mass (Sheehan, 2021). This results in an estimated carbon footprint of ~-1.45 kgCO₂e/kg ethanol, though a more detailed and thorough lifecycle analysis that includes cradle-to-grave considerations is the subject of a future study that is currently underway.

Especially when transportation GHG emissions are minimal as to keep the net CO₂e <100 kg per ton of CO₂ captured, CR is an ideal capture medium for downstream flow CO₂ conversion. The CO₂ is delivered as a liquid, which has a constant vapor pressure. This is helpful in flow systems if there are one or more stages of compression prior to introduction of CO₂ to other reactants. The suction pressure of these compressors must remain constant for optimum operation and to ensure adequate compressor lifetime, and severe variation in inlet CO₂ pressure or temperature can cause challenges that prompt plant shutdown. CR eliminates these operational variables and is economic without substantial subsidy, which makes it a model system for integration with flow CO₂ conversion systems in the near-term despite its limited long-term utility.

Unlike fermentation processes, NGCC power plants release flue gas streams with CO₂ concentrations of 4–16%. In this case, CO₂ is a harmful byproduct of electricity production. PSC from NGCC power plants takes place today using a commercial monoethanolamine (MEA) process, in which CO₂ is absorbed by liquid MEA at high pressure and low temperature and stripped from the MEA at low pressure and high temperature. Air Company's Calgary commercial CO₂-to-alcohols demonstration plant is deployed at the Shepard Energy Center, an 860 MW NGCC power plant. The NGCC byproduct flue gas containing 6.6% CO₂ is fed into the MEA adsorption system operated by the Alberta Carbon Conversion Technology Center (ACCTC), shown in the left on the photograph in Figure 2A. The product from the capture system, water-saturated CO₂ (98%), is then pumped into the Air Company building. Tail gas is typically emitted from amine CO₂ capture systems, which contains ~1.2% CO₂ that is not captured because it is too energy-intensive to do so. The water content of the captured CO₂ could have implications for the efficacy of the system, and a knockout drum dryer is used for its removal. After compression, the CO₂ is combined with H₂ and introduced into a reactor similar to, but significantly larger than, the Air Company pilot plant. While H₂ was supplied via tube trailer in the interim, construction of an integrated facility with a H₂O electrolyzer powered by renewable electricity is nearly complete. Table 4 shows a summary of the gas inlet parameters between the pilot and demonstration facilities.

For the Air Company pilot and commercial demonstration facilities as well as deployments for larger CO₂ utilization systems by us and others in the future, plant economics will play a major role in the technology used for CO₂ capture. Our experiences suggest that the hypothesis to target the high-concentration CO₂ emitters first is valid, but the GHG reduction of these sources is limited. This calls for research to reduce the capital costs, energy requirements, and improve product characteristics (e.g., temperature, pressure) for PSC and DAC at-scale. In our case, using a CR-sourced pilot reactor and PSC-sourced commercial demonstration reactor, the biggest barrier to use of DAC was the large capital expenditure for small units (on the order of 1–10 tons per day of CO₂). Innovations in materials science and sorbent materials that drive down the capital cost of small DAC units would be hugely beneficial for distributed deployment, especially in industries where customers are willing to pay a substantial premium for modular and distributed DAC, thus offsetting its comparably larger operational expense.

Beyond capital expenditure and operational considerations, the necessity of retrofits for certain PSC technologies could represent key barrier to PSC deployment and receives little academic attention (Koelbl et al., 2014). Due in part to the costs of retrofits and the abundance of non-retrofittable power plants, DAC has surprisingly been identified as the less expensive option compared to PSC in one third of NGCC plants. Further select cases (e.g., microalgae cultivation) also make DAC energetically competitive despite being the furthest from commercial availability (Mangram, 2012; Wilcox et al., 2017; Azarabadi and Lackner, 2020; Hirsch and Foust, 2020). There is no fundamental reason why both DAC and PSC systems also cannot deliver product CO₂ with the same consistency, pressures, and temperatures as CR to optimize integration with downstream flow reactors. Beyond facilitating retrofits, research that improves the feedstock input and CO₂ output tolerances of DAC and PSC technologies could further accelerate commercialization.

Locations for geological storage and measures to mitigate leakage also represent key barriers to scaling NET outside CR, PSC, and DAC technologies themselves that must be addressed (Koelbl et al., 2014; von Strandmann et al., 2019). Since CCS does not produce a physical, saleable end product, exploration into further profitable NET opportunities, mass production and innovative infrastructural development, financial incentives, and international policy will be necessary to reach emissions targets (Honegger and Reiner, 2018; Hirsch and Foust, 2020; Olfe-Kräutlein, 2020). The high startup costs associated with early CCU deployment may potentially be overcome by following successful disruptive innovation models in electric vehicles (EVs) and EV infrastructure that move from high-end to mass markets to scale pragmatically, for example, as Tesla has done (Chen and Perez, 2018).

Ultimately, widespread CO₂ capture and utilization will be needed to meet emissions targets, but these technologies alone will not save us. Real infrastructural change to facilitate an economic trajectory of CO₂ capture deployment is required. Similar to the way the hardware and software for EVs existed prior to its currently accelerating adoption due to cultural and political changes, these fundamental pieces exist for CO₂ capture. CO₂ utilization technologies now provide an additional incentive to build the required infrastructure on local and global scales.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

GP performed research on CO₂ capture technologies, compiled information in Tables 1, 2, and drafted the main body of the article. SS conceptualized the research, the facilities in Figures 1, 2, and supervised the work. Both authors wrote and revised the manuscript.

All authors are employees of Air Company.